BPEC: Collaborative Research: Creating Personalized Learning Pathways by Managing Cognitive Load

华盛顿大学与圣路易斯科学中心合作，将开发和评估一个计算机编程环境，使用个性化的学习途径，以更好地参与和支持年轻学习者。在正规和非正规教育环境中，解决问题的例子是一个重要的学习资源，使自学成为可能，并支持课堂上固有的异质性。最近对不同经验水平的程序员的研究表明，选择和调整在网络上共享的示例代码通常用于支持及时学习和访问不常用的技术。然而，尽管示例可以是强大的学习工具，但它们必须与学习者的经验水平和学习偏好相匹配。 该项目的工作人员将开发计算机算法，预测未来的问题和例子将最大限度地提高特定学习者的学习。通过例子学习的研究表明，通过选择和展示例子及其相关的实践问题来控制特定学习者的认知负荷，可以提高学习者在近距离和远距离迁移任务中的成功率。该建议假设，通过仔细控制认知负荷，将有可能构建个性化的学习路径，帮助学习者有效地从新手理解到掌握一个概念。本项目旨在回答以下问题：1）是否有可能预测感知的认知负荷为未来的问题，给定学习者的历史，并使用它来选择一个合适的下一个问题，为学习者？2)学习者历史中哪些因素最能预测未来问题的认知负荷？3)是否有可能有效地使用预测的认知负荷来构建个性化的学习路径为个人学习者？

Washington University, in collaboration with the St. Louis Science Center will develop and evaluate a computer programing environment that uses personalized learning pathways to better engage and support young learners. In both formal and informal education settings, examples of solved problems are an important learning resource, enabling self-teaching, and supporting the inherent heterogeneity found in classrooms. Recent studies of programmers at varying experience levels revealed that selecting and adapting example code shared on the web is often used to support just in time learning and to access infrequently used techniques. However, while examples can be powerful learning tools, they must be well-matched to the learner's experience level and learning preferences. The project staff will develop computer algorithms that predict what future problems and examples will maximize learning for a specific learner.Research on learning via examples has demonstrated that by controlling the cognitive load for a given learner through the selection and presentation of examples and their related practice problems, it is possible to improve learner's success on near and far transfer tasks. This proposal hypothesizes that by carefully controlling the cognitive load, it will be possible to construct personalized learning pathways that help a learner to efficiently move from a novice understanding to mastery of a concept. This project aims to answer the following questions:1) Is it possible to predict the perceived cognitive load for a future problem, given a learner's history and use this to select an appropriate next problem for that learner? 2) What are the factors of a learner's history that are most predictive of the perceived cognitive load for a future problem?3) Is it possible to effectively use the predicted cognitive load to construct personalized learning pathways for an individual learner?